Learning robot tactile sensing for object manipulation

Abstract

Abstract-Tactile sensing is a fundamental component of object manipulation and tool handling skills. With robots entering unstructured environments, tactile feedback also becomes an important ability for robot manipulation.In this work, we explore how a robot can learn to use tactile sensing in object manipulation tasks. We first address the problem of in-hand object localization and adapt three pose estimation algorithms from computer vision. Second, we employ dynamic motor primitives to learn robot movements from human demonstrations and record desired tactile signal trajectories. Then, we add tactile feedback to the control loop and apply relative entropy policy search to learn the parameters of the tactile coupling. Additionally, we show how the learning of tactile feedback can be performed more efficiently by reducing the dimensionality of the tactile information through spectral clustering and principal component analysis. Our approach is implemented on a real robot, which learns to perform a scraping task with a spatula in an altered environment.

Learning Sensor Feedback Models from Demonstrations via Phase-Modulated Neural Networks

“…Forcetorque sensing is also used in [1], where a hand-designed feedback gain matrix maps deviations from the expected force-torque measurements to the grasp plan adaptation. Previous work on robotic tactile-driven manipulation with tools has tried to learn feedback models to correct the position plans for handling uncertainty between tools and the environment, via reinforcement learning [5] or motor babbling [13]. In our work, we propose to bootstrap the learning of feedback model from human demonstrations.…”

“…The disparity S actual − S expected = ∆S can be used to drive corrections for adapting to the environmental changes causing the deviated sensor traces. Previous work [5], [18] uses reinforcement learning to learn these corrective behaviors, also in form of feedback models. However, learning a good feedback policy via trial-and-error from scratch is a very slow process.…”

“…We also compare the effect of separating versus embedding the feature representation learning with overall parameter optimization under phase modulation. Chebotar et al [5] used PCA for feature representation learning, which was separated from the phase-dependent parameter optimization using reinforcement learning. On the other hand, PMNN embeds feature learning together with the parameter optimization under phase modulation, into an integrated process.…”

“…Thus, the larger goal of this research is to equip Associative Skill Memories with a general feedback modulation learning framework, as depicted in the block diagram in Figure 1. Data driven approaches to learning such feedback models have been proposed [3], [4], [5] in the past. Here, we present a learning framework that improves such data-driven approaches in generality and experimental validation.…”

Learning feedback terms for reactive planning and control

“…This online modulation is achieved via coupling term functions that create a forcing term based on sensory information -thus creating a reactive controller. The potential of adding feedback terms to the DMP framework has already been shown in a variety of different scenarios, such as modulation for obstacle avoidance [2], [3], [4], [5] and adapting to force and tactile sensor feedback [6], [7]. These approaches have relied on extensive domain knowledge to design the form of the feedback term.…”